In [1]:
%load_ext autoreload
%autoreload 2

init and load image - visium tutorial¶

In [15]:
# import git
# import os 
# import getpass
# token = getpass.getpass()
In [16]:
# git.Git('/nfs/team205/ny1/ThymusSpatialAtlas/ImageSpot/').clone('https://nadavyayon:'+token+'@github.com/nadavyayon/jupyter-annotator.git')

you can use my env but do not modify it!!!! conda activate /nfs/team205/ny1/miniconda/miniconda3/envs/imagespot

In [9]:
import os
import numpy as np 
from bokeh.plotting import show, output_notebook
import pandas as pd
import socket
os.chdir('/nfs/team205/ny1/ThymusSpatialAtlas/ImageSpot/tissue-tag/')
import tissue_tag as tt
import matplotlib.pyplot as plt
import scanpy as sc
output_notebook()
host = '5011'
Loading BokehJS ...

Step 1 - Load image and define annotation names and colors¶

In [12]:
plt.figure(figsize=[12,12])

vis_path = '/nfs/team205/ny1/Visium/ThymusAtlasDataSet/set160222/Align_and_detect_output/V11J11-099_Z4-THY-FO3_A1_TA11556493/'
im,ppm,df = tt.read_visium(SpaceRanger_dir_path=vis_path,use_resolution='hires',res_in_ppm=0.1)
coordinates = np.vstack((df[5],df[4]))
plt.imshow(im,origin='lower')
plt.plot(coordinates[0,:],coordinates[1,:],'.')
plt.title( 'ppm - '+str(ppm))
Out[12]:
Text(0.5, 1.0, 'ppm - 0.1')

Step 2 - (optional) Predict regions from image¶

scribble iniital annotations¶

Step 2 - Manually draw annotations or Correct forest annotations (Step 2)¶

In [96]:
# define annotations Lv0 
annodict = {
    'lobes':'green'
}
# at this stage, if you are adding annotations make sure they come after the current ones otherwise it will mess up everything
result = np.zeros((im.shape[0],im.shape[1]), dtype=np.uint8) # blank annotation image

result_rgb = tt.rgb_from_labels(result,colors=list(annodict.values()))
out_img = tt.overlay_lebels(im,result_rgb,alpha=0.5,show=False)
tabs,render_dict = tt.poly_annotator(im,out_img,anno_dict=annodict,fig_downsize_factor=1)

# plotting
def app(doc):
    global tabs
    doc.add_root(tabs)
show(app,notebook_url=f'{socket.gethostname()}:'+host)
print('this takes up to a few minutes on farm')

this takes up to a few minutes on farm

plot annotations¶

In [100]:
result,object_dict  = tt.object_annotator(im,result,anno_dict=annodict,render_dict=render_dict,alpha=0.5)
print(object_dict)
# plot result 
result_rgb = tt.rgb_from_labels(result,colors=list(object_dict.values()))
out_img = tt.overlay_lebels(im,result_rgb,alpha=0.5)
plt.show()

lobes
{'unassigned': 'yellow', 'lobes_0': 'orange', 'lobes_1': 'blue', 'lobes_2': 'red', 'lobes_3': 'magenta'}

save annotation¶

In [101]:
isExist = os.path.exists(vis_path+'tissue_annotations')
if not(isExist):
    os.mkdir(vis_path+'/tissue_annotations/')
    
tt.save_annotation(
    folder=vis_path+'/tissue_annotations/',
    label_image=result,
    file_name='annotations_'+list(annodict.keys())[0],
    anno_names=list(object_dict.keys()),
    anno_colors=list(object_dict.values()),
    ppm = ppm,
)
In [102]:
# load tissue annotations from jupyter annotator 
annotations,annotations_labels,ppm,annodict = tt.load_annotation(vis_path+'/tissue_annotations/','annotations_lobes',load_colors=True)
im,ppm,positions = tt.read_visium(SpaceRanger_dir_path=vis_path+'/',use_resolution='hires5K',res_in_ppm=ppm)
result_rgb = tt.rgb_from_labels(annotations,colors=list(annodict.values()))
out_img = tt.overlay_lebels(im,result_rgb,alpha=0.5)
print(np.unique(annotations))

loaded annotation image - annotations_lobes size - (791, 777)
loaded annotations
{1: 'unassigned', 2: 'lobes_0', 3: 'lobes_1', 4: 'lobes_2', 5: 'lobes_3'}
loaded ppm
{'ppm': 0.1}
loaded color annotations
{'unassigned': 'yellow', 'lobes_0': 'orange', 'lobes_1': 'blue', 'lobes_2': 'red', 'lobes_3': 'magenta'}
[1 2 3 4 5]
Out[102]:
(791, 777, 4)

migrate annotations to hi res grid sapce¶

In [103]:
%%time
# it's critical not to split cell part and recalculate the df everytime you cahnge the dist2cluster resolution - I have an overflow bug somewhere that I need to fix
spot_diameter = 30 # don't go crazy here avoid under 15um spots 
df = tt.grid_anno(
    im,
    [annotations],
    ['annotations'],
    [annotations_labels],
    spot_diameter=spot_diameter,
    pixels_per_micron=ppm
)
df

generating grid with spot size - 30, with resolution of - 0.1 ppm
annotations

68640it [00:09, 7291.61it/s]

CPU times: user 9.45 s, sys: 70.5 ms, total: 9.52 s
Wall time: 9.47 s
Out[103]:
x y annotations annotations_number
index
0 3.0 3.000000 unassigned 1
1 3.0 5.984791 unassigned 1
2 3.0 8.969582 unassigned 1
3 3.0 11.954372 unassigned 1
4 3.0 14.939163 unassigned 1
... ... ... ... ...
68635 774.0 777.560852 unassigned 1
68636 774.0 780.545654 unassigned 1
68637 774.0 783.530396 unassigned 1
68638 774.0 786.515198 unassigned 1
68639 774.0 789.500000 unassigned 1

68640 rows × 4 columns

In [80]:
# remove background spots
# df = df.iloc[np.where(df['annotations']!='unassigned')[0],:]

move annotation and distances to visium spots¶

In [104]:
# calculate distance matrix between hires and visium spots
positions = tt.anno_to_visium_spots(
    df_spots=positions,
    df_grid=df,
    plot=False,
)
positions

make sure the coordinate systems are alligned e.g. axes are not flipped
migrating - annotations to segmentations
migrating - annotations_number to segmentations
Out[104]:
1 2 3 4 5 annotations annotations_number
0
GTATTCTTACCGTGCT-1 1.0 7.0 21.0 120.122387 178.557632 lobes_3 5
GCGGTAACCCAAATGA-1 1.0 7.0 23.0 120.143196 188.515124 lobes_3 5
CTACGTGTTGCCACCA-1 1.0 6.0 24.0 111.467050 193.499541 lobes_3 5
CTAGATAAACTCCTCG-1 1.0 7.0 25.0 120.164004 198.472617 lobes_3 5
TCCATTAGTTGGATAG-1 1.0 6.0 26.0 111.487859 203.457034 lobes_3 5
... ... ... ... ... ... ... ...
CTTGATGACCATCCAG-1 1.0 70.0 98.0 668.176154 561.563845 lobes_0 2
CTTGCCCACCCACGCA-1 1.0 70.0 100.0 668.196963 571.521337 lobes_0 2
GGACAACCATGAAGCC-1 1.0 70.0 102.0 668.217771 581.478830 lobes_0 2
AGGTGCACGTCCACAT-1 1.0 70.0 104.0 668.238580 591.436322 lobes_0 2
ACCAAGTCATCGGCAG-1 1.0 70.0 106.0 668.259389 601.393815 lobes_0 2

2958 rows × 7 columns

In [105]:
import scanpy as sc
adata_vis = sc.read_h5ad(vis_path+'/Corrected_AnnData.h5ad')
adata_vis.obs = pd.concat([adata_vis.obs,positions.iloc[:,5:]],axis=1)
sc.pl.spatial(adata_vis,color=['annotations'],cmap='gist_rainbow',ncols=2)

/nfs/team205/ny1/miniconda/miniconda3/envs/imagespot/lib/python3.9/site-packages/anndata/_core/anndata.py:1830: UserWarning: Variable names are not unique. To make them unique, call `.var_names_make_unique`.
  utils.warn_names_duplicates("var")
In [ ]: